J Appl Physiol 95: 1706–1716, 2003; 10.1152/japplphysiol.00288.2003. highlighted topics

Physiology of Aging Invited Review: Theories of aging

Brian T. Weinert and Poala S. Timiras Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, California 94720-3202

Weinert, Brian T. and Paola S. Timiras. Physiology of Aging. Invited Review: Theories of aging. J Appl Physiol 95: 1706–1716, 2003; 10.1152/japplphysiol.00288.2003.—Several factors (the lengthening of the average and, to a lesser extent, of the maximum human life span; the increase in percentage of elderly in the population and in the proportion of the national expenditure utilized by the elderly) have stimulated and continue to expand the study of aging. Recently, the view of aging as an extremely complex multifactorial process has replaced the earlier search for a distinct cause such as a single gene or the decline of a key body system. This minireview keeps in mind the multiplicity of mechanisms regulating aging; examines them at the molecular, cellular, and systemic levels; and explores the possibility of interactions at these three levels. The heterogeneity of the aging phenotype among individuals of the same species and differences in longevity among species underline the contri- bution of both genetic and environmental factors in shaping the life span. Thus, the presence of several trajectories of the life span, from incidence of disease and disability to absence of pathology and persistence of function, suggest that it is possible to experimentally (e.g., by calorie restriction) prolong functional plasticity and life span. In this minire- view, several theories are identified only briefly; a few (evolutionary, gene regulation, cellular senescence, free radical, and neuro-endocrine- immuno theories) are discussed in more detail, at molecular, cellular, and systemic levels. evolution; gene regulation; cellular senescence; free radical; neuro-endo- crine-immunologic regulation

IN RECENT DECADES, THE STUDY of aging has expanded normal aging process alone or in combination with rapidly both in depth and in breadth. This growth has other theories. The definition of aging itself is open to been stimulated by 1) the extraordinary lengthening of various interpretations (14, 79). In response to the the average human life span, worldwide; 2) the less question “Why do we age?” aging is presented as an spectacular, but nevertheless significant, lengthening ontogenic issue; the process of growing old and/or the of the maximum life span; 3) the increasing percentage sum of all changes, physiological, genetic, molecular, of elderly in the population, especially in some devel- that occur with the passage of time, from fertilization oped countries; and 4) the increased proportion of the to death. In response to the question “Why do we live as national health expenditures utilized by the elderly long as we do?” an evolutionary-comparative frame- (96). Biological, epidemiologic, and demographic data work is the preferred address. To the question “Why do have generated a number of theories that attempt to we die?” the answer should underline the lack of nec- identify a cause or process to explain aging and its essary relation between aging (a definite period of the inevitable consequence, death. However, in recent life span) and death (an event that may occur at all years, the search for a single cause of aging, such as a ages). However, because aging is characterized by the single gene or the decline of a key body system, has declining ability to respond to stress and by increasing been replaced by the view of aging as an extremely homeostatic imbalance and incidence of pathology, complex, multifactorial process (43). Several processes death remains the ultimate consequence of aging. The- may interact simultaneously and may operate at many ories formulated to explain aging processes have been levels of functional organization (31). Similarly, differ- grouped into several categories, some of the most ent theories of aging are not mutually exclusive and widely used being the programmed and error theories may adequately describe some or all features of the of aging. According to the “programmed” theories, ag- ing depends on biological clocks regulating the timeta- ble of the life span through the stages of growth, Address for reprint requests and other correspondence: P. S. Timiras, Dept. of Molecular and Cell Biology, 401 Barker Hall, development, maturity, and old age: this regulation Berkeley, CA 94720-3202 (E-mail: [email protected]). would depend on genes sequentially switching on and

1706 8750-7587/03 $5.00 Copyright © 2003 the American Physiological Society http://www.jap.org INVITED REVIEW 1707 off signals to the nervous, endocrine, and immune radical reduction), cellular (e.g., mitochondrial protec- systems responsible for maintenance of homeostasis tion), and systemic (e.g., endocrine shifts) mechanisms and for activation of defense responses. The “error” (57). Although these interventions extend beyond the theories identify environmental insults to living organ- limits of the theories of aging themselves, they will be isms that induce progressive damage at various levels mentioned here in their support. Some of the principal (e.g., mitochondrial DNA damage, oxygen radicals ac- theories of aging to be discussed here are listed in cumulation, cross-linking). Table 1: several will be identified only briefly, whereas In the present review, we have categorized the var- a few will be discussed in detail. The latter include ious theories of aging as evolutionary, molecular, cel- evolutionary, gene regulation, cellular senescence, free lular, and systemic. The choice of these categories and radical, and neuro-endocrine-immuno theories. the order in which they are presented reflect their affinity to physiological discourse (90). Thus theories of EVOLUTIONARY THEORIES aging may overlap at various levels of organization: alterations with aging of molecular events may lead to Why do we live as long as we do? Evolutionary cellular alterations, and these, in turn, contribute to theories argue that aging results from a decline in the organ and systemic failure with evolutionary implica- force of natural selection. Because evolution acts pri- tions for reproduction and survival. In complex, multi- marily to maximize reproductive fitness in an individ- cellular organisms, the study of interactions among ual, longevity is a trait to be selected only if it is intrinsic (genetic), extrinsic (environmental), and sto- beneficial for fitness. Life span is, therefore, the result chastic (random damage to vital molecules) causes of selective pressures and may have a large degree of provides a fruitful approach conducive to a comprehen- plasticity within an individual species, as well as sive and realistic understanding of the aging process. among species. The evolutionary theory was first for- In humans, for example, the current longevity is the mulated in the 1940s based on the observation that result of an early (middle of last century) “epidemio- Huntington’s disease, a dominant lethal mutation, re- logic transition,” referring to the decline in death rates mained in the population even though it should be due to acute infectious disease (because of improved strongly selected against (34). The late age of onset for hygiene and the discovery of antibiotics) (101). This Huntington’s disease (30–40 yr) allows a carrier to was followed in the 1970s to 1980s by a second mortal- reproduce before dying, thereby allowing the disease to ity decline at older ages in the reduction of death rates avoid the force of natural selection. This observation due to cardiovascular disease (101). In several animal inspired the Mutation Accumulation Theory of aging, species (rodents, monkeys), experimental interven- which suggests that detrimental, late-acting mutations tions such as restriction of dietary calories show that it may accumulate in the population and ultimately lead is possible to delay the onset of pathology and to pro- to pathology and senescence (59). Currently, there is long the life span by manipulating molecular (e.g., free scant experimental evidence for this theory of aging (67).

Table 1. Classification and brief description of main theories of aging

Biological Level/Theory Description Evolutionary Mutation accumulation* Mutations that affect health at older ages are not selected against. Disposable soma* Somatic cells are maintained only to ensure continued reproductive success; after reproduction, soma becomes disposable. Antagonistic pleiotropy* Genes beneficial at younger age become deleterious at older ages. Molecular Gene regulation* Aging is caused by changes in the expression of genes regulating both development and aging. Codon restriction Fidelity/accuracy of mRNA translation is impaired due to inability to decode codons in mRNA. Error catastrophe Decline in fidelity of gene expression with aging results in increased fraction of abnormal proteins. Somatic mutation Molecular damage accumulates, primarily to DNA/genetic material. Dysdifferentiation Gradual accumulation of random molecular damage impairs regulation of gene expression. Cellular Cellular senescence-Telomere theory* Phenotypes of aging are caused by an increase in frequency of senescent cells. Senescence may result from telomere loss (replicative senescence) or cell stress (cellular senescence). Free radical* Oxidative metabolism produces highly reactive free radicals that subsequently damage lipids, protein and DNA. Wear-and-tear Accumulation of normal injury. Apoptosis Programmed cell death from genetic events or genome crisis. System Neuroendocrine* Alterations in neuroendocrine control of homeostasis results in aging-related physiological changes. Immunologic* Decline of immune function with aging results in decreased incidence of infectious diseases but increased incidence of autoimmunity. Rate-of-living Assumes a fixed amount of metabolic potential for every living organism (live fast, die young). *Discussed in the text.

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org 1708 INVITED REVIEW

However, the basic concept that aging results from a insects have average life spans that are measured in lack of selection enjoys a wealth of experimental sup- months, not years. port. Long-lived Drosophila strains can be bred by selecting the offspring of older adults, demonstrating MOLECULAR THEORIES that life span can be altered directly by selective pres- The Gene Regulation Theory of aging proposes that sure (68, 75). Life span is species specific because it is senescence results from changes in gene expression largely a function of survivability and reproductive (38). Although it is clear that many genes show strategy in a competitive environment. Consequently, changes in expression with age (71, 98, 104), it is organisms that die primarily from predation and envi- unlikely that selection could act on genes that promote ronmental hazards will evolve a life span optimized for senescence directly (42). Rather, life span is influenced their own particular environment. This idea was tested by the selection of genes that promote longevity (see in a natural environment by comparing mainland opos- above). Recently, DNA microarrays have been used to sums that are subject to predation to a population of assay genome-wide transcriptional changes with age in opossums living on an island free of predators (4). The several model organisms (71, 74, 98, 104). Genome- Evolutionary Theory predicts that the protected island level analysis allows researchers to compile a tran- opossums would have the opportunity to evolve a scriptional fingerprint of “normal” aging. This data can longer life span, if it were beneficial to fitness. Indeed, be compared with interventions that slow or accelerate island opossums do live longer and age more slowly aging, perhaps enabling the identification of gene ex- than their mainland counterparts (4). The observation pression changes that are relevant to the aging process that organisms can age in a natural environment (51) (71, 99, 104). indicates that although extending life span can be One of the most exciting developments in aging re- beneficial to fitness, other considerations might neces- search is the identification of an insulin-like signaling sitate sacrificing longevity for reproductive fitness. pathway that regulates life span in worms, flies, and This basic idea of the Disposable Soma Theory of aging mice (87). Life span extension results from the activa- argues that the somatic organism is effectively main- tion of a conserved transcription factor in response to a tained only for reproductive success; afterward it is reduction in insulin-like signaling, indicating that disposable. Inherent in this theory is the idea that gene expression can regulate life span. Understanding somatic maintenance, in other words, longevity, has a how nature delays aging through changes in gene ex- cost; the balance of resources invested in longevity vs. pression should reveal much about the process of aging reproductive fitness determines the life span. itself and provide a starting point for developing ther- The concept of an evolutionary tradeoff is essential apies to slow aging. in both the Disposable Soma Theory and the Antago- Studies of human centenarians and their relatives nistic Pleiotropy Theory. The Disposable Soma Theory have identified a significant genetic aspect of the abil- explains why we live for a certain period of time but ity to survive to exceptional ages. In one study, the does not postulate the specific cause of aging, whereas mortality rate of centenarian siblings was shown to be, the theory of Antagonistic Pleiotropy suggests that on average, half the mortality rate of the U.S. year some genes may be selected for beneficial effects early 1900 cohort (69, 70). This sustained life-long reduction in life and yet have unselected deleterious effects with in mortality rate is taken to imply that the effect is due age, thereby contributing directly to senescence. An- to genetic rather than environmental or socioeconomic tagonism between reproduction and longevity is sup- factors. A recent study supports the idea that excep- ported by experiments in which limiting reproduction tional longevity has a genetic component by identifying by destroying germ line cells can extend life span in a locus on chromosome 4 that may contain gene(s) that both Drosophila and Caenorhabditis elegans (1, 83). In promote longevity (72). Genetic analysis of human lon- humans, the growth and normal function of the pros- gevity is especially important given that genetic as- tate gland are promoted by androgens, the male go- pects of aging are studied primarily in short-lived nadal hormones. In old age these same hormones may model organisms. It will be particularly interesting to contribute to the etiology of prostate cancer, one of the see whether the recent advances in understanding major causes of death in old men. The relationship genetics of longevity in model organisms are verified in between longevity and fecundity is not absolute; some human studies, and vice versa. long-lived Drosophila strains have no loss in reproduc- tive capacity (2), and long-lived three-toed box turtles CELLULAR THEORIES continue to reproduce for more than 60 yr (64). Ani- Cell senescence/telomere theory. The Cellular Senes- mals (such as the box turtle above) that adapt to escape cence Theory of aging was formulated in 1965 when predation might favor the selection of both longevity cell senescence was described as the process that limits and fecundity. For example, eusocial insects, such as the number of cell divisions normal human cells can ants, will cooperate to support a single queen. The undergo in culture (36). This “limit in replicative ca- queen, protected from the environment and cared for pacity” occurs after a characteristic number of cell by worker ants, will give rise to hundreds and even divisions and results in terminally arrested cells with thousands of offspring each day and, in some cases, altered physiology (11). Cellular senescence can also lives for 30 years (40). In contrast, related, noneusocial occur in response to distinct molecular events; in this

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org INVITED REVIEW 1709 discussion, we distinguish between cellular senescence The tumor suppressor protein p53 is a key regulator (of all types) from senescence due to cell replication of cellular checkpoint responses to genome crisis. (replicative senescence) and senescence due to other Among the many functions attributed to p53 are essen- causes [stress-induced senescence (SIS)]. Replicative tial roles in activating transient cell cycle arrest, apop- senescence is a specific type of cellular senescence that tosis, replicative senescence, and SIS in response to ultimately results from loss of telomeres (specialized radiation-induced DNA damage and replication-in- structures composed of a repeating DNA sequence and duced telomere loss (23). The type of p53-dependent located at the ends of each linear chromosome; Ref. 8). cellular response (cell arrest, apoptosis, or senescence) With each cell division, a small amount of DNA is is often dependent on the particular cell type being necessarily lost at each chromosome end, resulting in examined or the type and severity of stress that the ever-shorter telomeres, altered telomere structure, and cells are exposed to. Mice mutated for p53 have a eventual replicative senescence (8). Activation of the dramatically increased incidence of cancer (24), and telomerase enzyme will regenerate telomeres, prevent p53 signaling is altered in ϳ80% of human cancers replicative senescence, and immortalize human pri- (23), indicating that p53-mediated functions are impor- mary cell cultures (10). SIS occurs in response to a tant for tumor suppression. Replicative senescence variety of stressors, including but not limited to 1) and/or SIS may have the biological role of preventing DNA damage, 2) modifications in heterochromatin cancer by limiting the replicative potential of any given structure, and 3) strong mitogenic signals resulting cell. However, if cellular senescence acts to suppress from oncogene expression (11). Specialized immortal tumor formation, then how do we explain the observa- cell types such as stem cells, germ cells, and T lympho- tion that both cancer and cellular senescence are more cytes express telomerase and will either maintain telo- prevalent with age? One way in which this apparent mere length or delay telomere attrition (17, 102). Ad- contradiction can be resolved is by the evolutionary ditionally, all cancer cells activate telomerase or an hypothesis (antagonistic pleiotropy) that cellular se- alternate pathway of telomere extension to avoid rep- nescence was selected to suppress cancer early in life licative senescence (41, 73). yet has the unselected effect of contributing to age- Initial experiments with cells in culture showed a related pathology and cancer in older, postreproductive correlation between replicative potential and donor individuals (44). age, suggesting that cells from older individuals have a The requirement of telomerase for human cell im- more limited capacity for further cell divisions. Simi- mortality together with the observation that telomeres larly, organisms with short life spans have cells that shorten with age led to the speculation that telomere senesce more rapidly than organisms with longer life length regulates cell replicative life span in vivo and spans. However, recent experiments have cast consid- contributes to aging. Although difficult to test directly erable doubt on these observations, and further re- in humans, experiments in rodents have provided little search is required to clarify these divergent data (re- support for this idea. Gene targeting showed that te- viewed in Refs. 8, 78, 103). Cells expressing stress- lomerase-deficient mice do not age rapidly; in fact, induced markers found in senescent cells accumulate overt phenotypes are not observed for several genera- with age in many tissues (20, 44), although it remains tions (9, 49). This showed that telomere shortening unclear whether this indicates the presence of senes- cannot account for normal aging in mice; however, cent cells in vivo. Several studies suggest that athero- similarities between aging and the late-generation te- sclerosis results from senescent changes in arterial lomerase-deficient phenotype might indicate that cel- endothelial cells (15, 26, 94). Werner’s syndrome is a lular senescence of some type contributes to aging in remarkable progeroid syndrome due to an apparently mice. The tumor suppressor protein p53 is required for normal period of development until , followed cellular senescence; p53 deficiency suppresses the by early manifestation of many aging-related physio- early aging phenotype of late-generation telomerase- logical changes. Notable among these changes is the deficient mice (16). These data suggest that p53-depen- early onset of atherosclerosis (54); in addition, cells dent processes (including, but not limited to cellular from both Werner’s patients and a mouse model for the senescence) are responsible for the early aging pheno- disease are marked by accelerated senescence in cell type in telomerase-deficient mice, an interpretation culture (47, 55). The altered physiology of senescent supported by the recent finding that a hyperactive p53 cells might contribute to aging and cancer through mutant mouse ages rapidly and has a markedly re- secondary effects on neighboring cells in tissues (44). duced incidence of spontaneous tumors (92). The essen- For example, senescent endothelial cells upregulate tial role of p53 in cellular senescence is underscored by the proinflammatory cytokine interleukin-1␣ and recent reports indicating that p53 is required for main- EGF-like growth factors (50, 52). Such changes may tenance of cellular senescence in some cell types. result in a hazardous local environment in which in- Treatments that inactivate p53 in senescent cells can flammation and mitogenic stimulation lead to a decline trigger reentry into the replication cycle and cell pro- in organ function and increased risk of cancer. Consis- liferation (5a, 21), although some human senescent tent with this idea, the growth of preneoplastic and cells (those with elevated p16 expression) are refrac- neoplastic epithelial cells in culture is stimulated by tory to senescence release by p53 inhibition (5a). the presence of senescent fibroblasts compared with Although telomere shortening does not appear to presenescent fibroblasts (45). have a significant role in aging mice, there is some

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org 1710 INVITED REVIEW evidence that telomeres may contribute to normal hu- data suggest that, in mice, oxidative damage is respon- man aging. Dyskeratosis congenita (DKC) is an X- sible for cellular senescence in culture and may ac- linked disorder marked by skin and pa- count for cellular senescence in vivo, an interpretation thologies largely attributed to the loss of functional that lends credence to both free radical and cellular stem cells in these tissues (22, 25). The mutation re- senescence theories of aging. It is worth noting that the sponsible for DKC affects an enzyme involved in the question of replicative senescence vs. SIS has a wider metabolism of the telomerase RNA subunit (hTR) (65). implication in terms of theories of aging in general. A rare dominant autosomal form of DKC can result Replicative senescence can be considered a cause of from mutation of the hTR gene directly (95), support- aging in and of itself, as it is largely attributed to the ing the idea that DKC manifests itself through telom- number of cell divisions as determined by telomere erase dysfunction. Interestingly, patients with the length. On the other hand, SIS is a response to stress, dominant hTR-defective form of DKC have a more particularly genome crisis and DNA damage. SIS severe pathology in later generations (95), mirroring should, therefore, be considered a cellular response to the delayed phenotype observed in telomerase-defi- age-related molecular changes that likely acts to exac- cient mice (see above). Although DKC patients develop erbate or accelerate organismal aging. This view of pathologies that partly resemble normal aging, these cellular senescence in aging is compatible with the phenotypes are limited compared with a more exten- various damage accumulation theories (such as free sive progeroid disorder such as Werner’s syndrome and radical, error catastrophe, and somatic mutation) that suggest a limited role for telomere shortening in nor- may account for the ultimate cause of cellular senes- mal human aging. For example, telomere length may cence with aging. restrict the replicative potential of hemopoietic cells, Free radical theory. The Free Radical Theory of ag- perhaps contributing to the well-documented decline in ing was first proposed in 1957 (35); it is one of the immune function with age. Patients with DKC are not best-known theories and remains controversial to this completely telomerase deficient; depending on the spe- day. All organisms live in an environment that con- cific type of disease (X-linked or autosomal), telomer- tains free radical-containing reactive oxygen species ase levels may be from two- to fivefold reduced (18). (ROS); mitochondrial respiration, the basis of energy Interestingly, the age of disease onset may be corre- production in all eukaryotes, generates ROS by leaking lated with the degree of telomerase deficiency, with the intermediates from the electron transport chain (29). most deficient individuals developing pathologies at an The universal nature of oxidative free radicals is un- earlier age. An interesting model suggests that telo- derscored by the presence of superoxide dismutase mere shortening can promote tumorgenesis by enhanc- (SOD), an enzyme found in all aerobic organisms that ing genome instability: telomere-induced genome crisis scavenges superoxide anions exclusively (29). In addi- leads to cell transformation, which is followed by te- lomerase activation to allow for unlimited cell prolifer- tion, cellular oxidative damage is indiscriminate; there ation (56). Consistently, some DKC patients have an is evidence for the oxidative modification of DNA, pro- increased incidence of carcinomas, suggesting that tein, and lipid molecules (60). The Free Radical Theory telomere shortening may contribute to the develop- supposes that free radical reactivity is inherent in ment cancer that is more prevalent with age (18). biology and results in cumulative damage and senes- There is an ongoing debate as to the relative contri- cence. In fact, elevated levels of both oxidant-damaged butions of replicative senescence (due to telomere loss) DNA and protein are found in aged organisms (6, 86). and SIS (due to damage accumulation and other fac- Although it is clear that oxidative damage accumulates tors) in the aging process. The validity of conclusions with aging, it is not clear whether this process contrib- based on the replicative life span of cells in culture has utes to aging in all organisms. A more thorough review been questioned in several recent reviews (8, 78, 103). of the Free Radical Theory may be found in several In addition, experiments in mice have provided little, if excellent reviews that focus exclusively on this topic any, support for a role of replicative senescence in (29, 60). aging, although it is not unreasonable to assume that Some of the strongest evidence in support of the Free humans and mice may differ as to the ultimate causes Radical Theory comes from life span experiments in of cell senescence in culture (84). Recent results illus- flies and worms. The increased life span of transgenic trate this point by showing that mouse embryonic flies expressing SOD (91) indicates that free radical- fibroblasts (MEFs) enter SIS in response to the ele- scavenging enzymes are sufficient to delay aging in vated oxygen (20%) present in normal tissue culture, a Drosophila. In addition, flies selected for increased characteristic that distinguishes these cells from hu- longevity have elevated levels of SOD and increased man cells (66a). MEFs normally enter cellular senes- resistance to oxidative stress (3). Long-lived mutant cence after just 10–15 divisions in cell culture and with worms are also resistant to oxidative stress and show very long telomeres; this phenomenon was previously an age-dependent increase in SOD and catalase activ- considered the replicative life span of these cells. Cells ity (46). Extension of C. elegans life span by synthetic grown in 3% oxygen do not senesce at all, indicating small molecules that mimic catalase and/or SOD dem- that previous estimates of MEF replicative potential onstrates that antioxidant compounds can delay aging were based on observations of cells that enter SIS in worms (61). Clearly, free radical damage opposes owing to oxidative damage in tissue culture. These longevity in these small, short-lived organisms; but

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org INVITED REVIEW 1711 what about larger, long-lived organisms such as mam- SYSTEM-BASED THEORIES OF AGING: mals? NEUROENDOCRINE AND IMMUNE THEORIES Dietary antioxidants can reduce the accumulation of In these theories, the aging process is related to the oxidized molecules in mice, yet they fail to extend life decline of the organ systems essential for 1) the control span (60). Rodents with SOD mutations are quite sick and maintenance of other systems within an organism, and die prematurely, although it is not clear that they and 2) the ability of organisms to communicate and actually age rapidly. Ubiquitous overexpression of adapt to the environment in which they live. In hu- SOD does not extend life span in rodents, indicating mans, all systems may be considered indispensable for that SOD does not limit longevity (37). Ionizing radia- survival. However, the nervous, endocrine, and im- tion generates free radicals; surprisingly, chronic radi- mune systems play a key role by their ubiquitous ation exposure actually causes a reproducible increase actions in coordinating all other systems and in their in rodent life span (13). The longevity-enhancing effect interactive and defensive responsiveness to external of chronic radiation may be explained if radiation ex- and internal stimuli. posure results in stable activation of cellular defenses. Neuroendocrine theory. This theory proposes that Similar stress conditioning can lead a positive compen- aging is due to changes in neural and endocrine func- satory response (hormesis) that protects against oxida- tions that are crucial for 1) coordinating communica- tive damage and extends life span (29). Calorie restric- tion and responsiveness of all body systems with the tion is an intervention that prolongs the life span of external environment; 2) programming physiological nearly every organism to which it has been applied (see responses to environmental stimuli; and 3) maintain- below). In rodents, calorie restriction reduces genera- ing an optimal functional state for reproduction and tion of ROS from isolated mitochondrial preparations survival while responding to environmental demands. and attenuates the accumulation of oxidative damage These changes, often detrimental in nature, not only (63). Free Radical Theory may provide an attractive selectively affect the neurons and hormones that reg- explanation for the longevity-promoting effects of cal- ulate evolutionarily significant functions such as re- orie restriction (i.e., reducing dietary intake reduces production, growth, and development, but also affect metabolism and ROS production); however, calorie re- those that regulate survival through adaptation to striction is known to alter the function of many other stress. Thus the life span, as one of the cyclic body molecular, cellular, and organ systems (see below). functions regulated by “biological clocks,” would un- Although it is easy to find correlations between many dergo a continuum of sequential stages driven by ner- physiological functions and calorie restriction, it re- vous and endocrine signals. Alterations of the biologi- mains difficult to distinguish the ultimate cause of life cal clock, e.g., decreased responsiveness to the stimuli span extension by this technique from the abundant driving the clock or excessive or insufficient coordina- molecular and cellular changes that accompany it. tion of responses, would disrupt the clock and the The Free Radical Theory is further divided into sev- corresponding adjustments (27, 28, 88, 89). An impor- eral hypotheses focusing on the exclusive role of par- tant component of this theory is the perception of the ticular organelles and types of damaged molecules in hypothalamo-pituitary-adrenal (HPA) axis as the mas- the aging process. One such hypothesis argues that ter regulator, the “pacemaker” that signals the onset mutations in mitochondrial DNA accelerates free rad- and termination of each life stage. One of the major ical damage by introducing altered enzyme compo- functions of the HPA axis is to muster the physiological nents into the electron transport chain. Faulty electron adjustments necessary for preservation and mainte- transport results in elevated free radical leakage and nance of the internal “homeostasis” (steady state) de- ultimately more mitochondrial DNA mutation and ex- spite the continuing changes in the environment (7, acerbated oxidant production. This “vicious cycle” of 12). During life span, chronic exposure to severe stress mutation and oxidant production eventually leads to from a multitude of physical, biological, or emotional cellular catastrophe, organ failure, and senescence stimuli may exhaust or weaken the capacity to adapt (53). Another hypothesis argues that free radicals and lead to the so-called “diseases of adaptation” and cause aging when oxidized proteins accumulate in death (58, 82). Aging would then result from “a de- cells. An age-dependent reduction in the ability to creasing ability to survive stress,” one of the many degrade oxidatively modified proteins may contribute definitions of aging that suggests a close relationship to the build-up of damaged, dysfunctional molecules in between stress and longevity. the cell (86). The Somatic Mutation Theory of aging Integration of responses to environmental stimuli supposes that accumulation of genetic mutations in would be carried out by the hypothalamus from infor- somatic cells is the specific cause of senescence; free mation derived in various cerebral structures (primar- radical damage may be an important source of somatic ily the cerebral cortex, limbic lobe, and reticular for- mutations (6). However, mice have been serially cloned mation). The hypothalamus itself regulates 1) several by somatic nuclear transfer for six generations without important nervous functions (e.g., sympathetic and any sign of premature aging (97), indicating that so- parasympathetic visceral functions), 2) behaviors (e.g., matic mutations cannot account for aging in mice and sexual and eating behaviors, rage, fear), and 3) endo- free radicals are not likely to promote senescence in crine functions, such as producing and secreting hy- this manner. pophysiotropic hormones that stimulate or inhibit hor-

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org 1712 INVITED REVIEW mone release from the pituitary (or hypophysis). In grated responses play an important role in monitoring response to hypothalamic signals, the pituitary, often metabolic and reproductive status to permit appropri- referred to as the master endocrine gland, produces ate energy adjustments and, ultimately, extend life and secretes several hormones that act to regulate span (87). Thus it may be assumed that this primitive many important functions of the body. Pituitary regu- neuroendocrine system has the capacity not only to lation occurs by releasing hormones (e.g., growth hor- coordinate what occurs in each cell and tissue of the mone, oxytocin, vasopressin), or by stimulating a pe- body, but also to avoid disorganization (e.g., overex- ripheral endocrine gland such as the adrenal cortex, pression leading to toxicity) of stress responses. These thyroid, or gonads. The adrenal gland is formed of a landmark studies in nematodes encourage further ex- cortex that surrounds a central core or medulla. Major ploration of hierarchical programming among the mul- hormones of the adrenal medulla are the cat- tiple factors that regulate longevity. echolamines epinephrine and norepinephrine, which Neuroendocrine-immuno theory. In the hierarchy of function as neurotransmitters for the sympathetic di- multisystem regulation throughout the sequential vision of the autonomic nervous system: these respond stages of life, there is a significant role for the interac- rapidly to any external or internal stress through cir- tion and integration of the neuroendocrine and im- culatory (increased blood pressure) and metabolic (fa- mune systems. Such interaction occurs through 1) neu- cilitating carbohydrate and lipid utilization for energy) ropeptides and cytokines present in the immune sys- adjustments (12). With aging, a reduction in sympa- tem that mediate both intraimmune communication thetic responsiveness is characterized by 1) a de- and communication between the neuroendocrine and creased number of catecholamine receptors in periph- immune systems, 2) several hormones from the poste- eral target tissues; 2) a decline of heat shock proteins rior (vasopressin) and anterior (thyroid-stimulating that increase stress resistance in many animal species, hormone, prolactin, adrenocorticotropic hormone, and including humans, and 3) a decreased capability of ) pituitary that control many impor- catecholamines to induce these heat shock proteins tant immune functions (cytokine and antibody produc- (93). The hormones of the adrenal cortex are glucocor- tion, lymphocyte cytotoxicity and proliferation, and ticoids, for the regulation of lipid, protein, and carbo- macrophage function), and 3) reciprocal action of cyto- hydrate metabolism; mineralocorticoids, for that of wa- kines on neuroendocrine functions. For example, inter- ter and electrolytes; and sex hormones. Among the leukin-1 activates the HPA by stimulating the secre- latter is , which decreases tion of cortico-releasing and adrenocorticotropic hor- with aging; dehydroepiandrosterone replacement ther- mones and may also act on the release of other apy has been advocated in humans, despite unconvinc- pituitary hormones (thyroid-stimulating hormone, ing results (19). Glucocorticoids, as well as other (ovar- growth hormone, prolactin, luteinizing hormone). ian and testicular) steroid hormones, are regulated by Parallel to neuroendocrine interactions, the immune positive and negative feedback between the target hor- system has several essential functions. The immune mones and their central control by the pituitary and system must control and eliminate foreign organisms hypothalamus. With aging and in response to continu- and substances in the host body while at the same time ing and severe stress, not only feedback mechanisms recognizing and therefore sparing from destruction the may be impaired, but also glucocorticoids themselves molecules (cells and tissues) from oneself. In most may become toxic to neural cells, thus disrupting feed- elderly humans, is characterized back control and hormonal cyclicity (80, 81). by a decreased resistance to infectious diseases, a de- The Neuroendocrine Theory has recently been sup- creased protection against cancer, and an increased ported by data showing that an “ancestral” insulin failure to recognize self (hence, autoimmune pathol- pathway controls stress responses and longevity in the ogy) (31, 33). However, different immune responses are nematode C. elegans (see also above) (39). Mutations of differentially affected with age. In humans, the a number of genes in this pathway confer 1) resistance is one of the most important immune organs: it is to environmental stress, including heat shock (93), 2) involved in the selection and maturation of T cells and enhanced resistance to starvation, and 3) extended production of peptide hormones. The thymus reaches a longevity. Many of these same genes are conserved in peak in both size and function during puberty; shortly humans: the insulin/insulin-like growth factor-I thereafter, it atrophies and progressively reduces its (IGF-I) peptide and Daf-2 gene are homologs of the production of mature T cells and hormones. This sign human insulin and IGF-I receptor, unc-64 and unc-31 of early immunosenescence may be interpreted as a are homologous to human synthaxine and catabolite tradeoff between the decreasing usefulness of the thy- activator proteins that are involved in the release of mus once the repertoire of T cells has been set up and neurotransmitters at the synapse, Age-1 is related to a the cost of maintenance of the organ. (32). Yet other conserved phosphoinositol-3-kinase that responds to functions, for example the activities of several types of insulin receptor activation, and Daf-16 is the homolog lymphocytes (natural killer and dendritic cells, macro- of the human forkhead box, class-O transcription factor phages) and of the complement system, are well pre- (5). In C. elegans, a relatively complex organism, these served in healthy centenarians (30). genes constitute a primordial neuroendocrine system Both the neuroendocrine and immune systems are in which the insulin/IGF-I peptide integrates informa- endowed with a high degree of plasticity, that is, the tion from environmental stress. The resulting inte- ability to modify their function according to demand.

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org INVITED REVIEW 1713

Plasticity is most efficient at early ages but also per- were diminished (48). Caloric restriction may act to sists at advanced age. Although studies of human ag- promote longevity through metabolic reprogramming ing in the 1960s to 1980s focused on functional decre- with a transcriptional shift (perhaps triggered by in- ments with aging involving all organs and systems of sulin) toward reduced energy metabolism and in- the body (85), in the 1980s and 1990s there was a creased biosynthesis and turnover of proteins. Caloric reconceptualization of the aging process that 1) deem- restriction also markedly influences the expression of phasized the view that aging is exclusively character- pathological phenotypes in rodent species selectively ized by declines in function and in health, 2) refocused bred as models of human pathology. However, the on the substantial heterogeneity among older persons, many benefits of caloric restriction are accompanied by 3) underscored the existence of positive trajectories of a number of untoward effects that may prevent its aging (i.e., without disease, disability, and major phys- applicability in humans and other animals; among iological decline), and 4) highlighted the possible avoid- these, the most significant are delayed (or stunted) ance of many usually aging-related diseases and dis- growth and failure of sexual maturation. The molecu- abilities (76). Thus three trajectories of aging were lar mechanisms of caloric restriction remain unre- delineated, the first characterized by disease and dis- solved; however, this intervention is a useful experi- ability, the second, known as “usual aging,” character- mental manipulation of aging in a variety of animal ized by absence of overt pathology but presence of some species, a property that fully merits its current wide- decline in function, and the last, the so-called “success- spread use in the study of aging. ful aging,” with little or no physiological loss and no pathology (76). Mechanisms of successful aging would CONCLUDING REMARKS consist of 1) persistence of normal function and plas- ticity, 2) compensatory responses to restore normal It should be clear from the content of this review that function (as may be induced by exercise, good nutri- the ultimate causes of aging remain unknown. On the tion, and better education), 3) interventions to replace other hand, a great deal of the aging process is under- deficient function (as represented by replacement ther- stood and may only require the integration of various apies), 4) changing of health outcome by modifying risk models and theories to account for normal aging. In our profiles (as in Ref. 2), 5) prevention of disease, and 6) view, the aging process is multifactorial and complex, strengthening of social interactions and support (77). A but not irreducibly so. Many of the pleiotropic changes successful example of this “functional remodeling” may that occur with aging may result from one or more be mediated by neuroendocrine and immune signals primary changes that affect many downstream pro- (66). For example, insulin sensitivity by peripheral cesses. This interconnectivity of the aging process often target tissue is decreased in old age but may be im- obfuscates the root cause of aging and limits the ability proved through caloric restriction (100). Another exam- to draw definitive conclusions from experimental re- ple is the significant lengthening (by 40% and more) of sults. Life span extension by calorie restriction is often the life span induced by caloric restriction (57). This cited in support of one or another theory of aging. For experimental intervention acts at various levels of example, calorie restriction reduces oxidant production function and involves a number of molecular, cellular, from mitochondria (see above), and it also prevents or and systemic changes. Only a few aspects of caloric delays age-related changes in endocrine function (such restriction will be discussed below. as estrogen receptor density in the hypothalamus). A Caloric restriction is the most potent and reproduc- free radical theorist may argue that oxidative damage ible environmental variable capable of extending the causes aging in a universal manner and that the life span in a variety of animals from worms to rats. changes in endocrine function with calorie restriction This simple intervention is achieved by providing a are secondary to changes in oxidant production in diet containing all the essential nutrients and vitamins endocrine cells. An endocrinologist might offer the but significantly restricted (by 30–70%) in calories. In counterargument that, because hormones regulate me- addition to the severity of the restriction, the degree of tabolism, calorie restriction delays aging by acting on life span lengthening depends on several factors: the the endocrine system directly, and all other physiolog- specific animal species, age at onset of restriction, and ical changes are secondary to this effect. One of the others (62). Not only is longevity increased but also most important goals in aging research is to determine metabolic responses (e.g., increased tissue sensitivity how a physiological intervention such as calorie re- to insulin), neuroendocrine and immune responses striction signals the body to delay aging. Is it a passive (e.g., increased defenses against stress, , can- process dependent on metabolic changes that accom- cer), and collagen responses (e.g., reduction of cross- pany reduced caloric intake, or is the organism actively linking) are significantly enhanced (66). Such func- responding to a caloric reduction to prolong reproduc- tional changes may be associated with changes in gene tive life span? At present, the answer is not entirely expression profile. For example, after chronic (28 mo) clear. The ability of insulin-like signaling to regulate severe (76%) caloric restriction, the genetic changes life span argues for the latter, even though a definitive that occurred in aging mice fed ad libitum (i.e., non- connection between calorie restriction and insulin-like caloric restricted) were significantly (by 84%) attenu- signaling awaits demonstration. However, the ability ated: for those genes characterized by increased ex- to study an active regulatory system that affects life pression, 29% were completely prevented and 34% span is an enormous benefit to aging research, because

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org 1714 INVITED REVIEW we can now explore the molecular mechanisms that al. A biomarker that identifies senescent human cells in culture connect changes in gene expression due to insulin and in aging skin in vivo. Proc Natl Acad Sci USA 92: 9363– 9367, 1995. signaling (and perhaps calorie restriction) with its ul- 21. Dirac AM and Bernards R. Reversal of senescence in mouse timate consequence, the delay of aging. fibroblasts through lentiviral suppression of p53. J Biol Chem 278: 11731–11734, 2003. REFERENCES 22. Dokal I. Dyskeratosis congenita: recent advances and future 1. Arantes-Oliveira N, Apfeld J, Dillin A, and Kenyon C. directions. J Pediatr Hematol Oncol 21: 344–350, 1999. Regulation of life-span by germ-line stem cells in Caenorhabdi- 23. Donehower LA. Does p53 affect organismal aging? J Cell tis elegans. Science 295: 502–505, 2002. Physiol 192: 23–33, 2002. 2. Arking R, Buck S, Novoseltev VN, Hwangbo DS, and Lane 24. Donehower LA, Harvey M, Slagle BL, McArthur MJ, M. Genomic plasticity, energy allocations, and the extended Montgomery CA Jr, Butel JS, and Bradley A. Mice defi- longevity phenotypes of Drosophila. Ageing Res Rev 1: 209–228, cient for p53 are developmentally normal but susceptible to 2002. spontaneous tumours. Nature 356: 215–221, 1992. 3. Arking R, Burde V, Graves K, Hari R, Feldman E, Zeevi A, 25. Drachtman RA and Alter BP. Dyskeratosis congenita. Der- Soliman S, Saraiya A, Buck S, Vettraino J, Sathrasala K, matol Clin 13: 33–39, 1995. Wehr N, and Levine RL. Forward and reverse selection for 26. Fenton M, Barker S, Kurz DJ, and Erusalimsky JD. Cel- longevity in Drosophila is characterized by alteration of antiox- lular senescence after single and repeated balloon catheter idant gene expression and oxidative damage patterns. Exp denudations of rabbit carotid arteries. Arterioscler Thromb Gerontol 35: 167–185, 2000. Vasc Biol 21: 220–226, 2001. 4. Austad SN. Retarded senescence in an insular population of 27. Finch CE. Longevity, Senescence and the Genome. Chicago, IL: opossums. J Zool 229: 695–708, 1993. Univ. of Chicago Press, 1990. 5. Babar P, Adamson C, Walker GA, Walker DW, and Lith- 28. Finch CE. The regulation of physiological changes during gow GJ. P13-kinase inhibition induces dauer formation, ther- mammalian aging. Q Rev Biol 51: 49–83, 1976. motolerance and longevity in C. elegans. Neurobiol Aging 20: 29. Finkel T and Holbrook NJ. Oxidants, oxidative stress and 513–519, 1999. the biology of ageing. Nature 408: 239–247, 2000. 5a.Beausejour CM, Krtolica A, Galimi F, Nerita M, Lowe S, 30. Franceschi C, Monti D, Sansoni P, and Cossarizza A. The Yaswen Y, and Campisi J. Reversal of human cellular senes- of exceptional individuals: the lesson of centenar- cence: roles of the p53 and p16 pathways. EMBO J. In press. ians. Immunol Today 16: 12–16, 1995. 6. Beckman KB and Ames BN. The free radical theory of aging 31. Franceschi C, Valensin S, Bonafe M, Paolisso G, Yashin matures. Physiol Rev 78: 547–581, 1998. AI, Monti D, and De Benedictis G. The network and the 7. Bernard C. Lec¸ons sur les phe´nome`nes de la vie communs aux remodeling theories of aging: historical background and new animaux et aux ve´ge´taux. Paris: Bailliere, 1878–1879. perspectives. Exp Gerontol 35: 879–896, 2000. 8. Blackburn EH. Telomere states and cell fates. Nature 408: 32. George AJ and Ritter MA. Thymic with ageing: 53–56, 2000. obsolescence or good housekeeping? Immunol Today 17: 267– 9. Blasco MA, Lee HW, Hande MP, Samper E, Lansdorp PM, 272, 1996. DePinho RA, and Greider CW. Telomere shortening and 33. Ginaldi L and Sternberg H. The immune system. In: Physi- tumor formation by mouse cells lacking telomerase RNA. Cell ological Basis of Aging and Geriatrics (3rd ed.), edited by 91: 25–34, 1997. Timiras PS. Boca Raton, FL: CRC, 2003. 10. Bodnar AG, Ouellette M, Frolkis M, Holt SE, Chiu CP, 34. Haldane JBS. New Paths in Genetics. London: Allen & Unwin, Morin GB, Harley CB, Shay JW, Lichtsteiner S, and 1941. Wright WE. Extension of life-span by introduction of telomer- 35. Harman D. Aging: a theory based on free radical and radiation ase into normal human cells. Science 279: 349–352, 1998. chemistry. J Gerontol 2: 298–300, 1957. 11. Campisi J. Cellular senescence and cell death. In: Physiolog- 36. Hayflick L. The limited in vitro lifetime of human diploid cell ical Basis of Aging and Geriatrics (3rd ed.), edited by Timiras strains. Exp Cell Res 37: 614–636, 1965. PS. Boca Raton, FL: CRC, 2003, p. 47–59. 37. Huang TT, Carlson EJ, Gillespie AM, Shi Y, and Epstein 12. Cannon WB. The Wisdom of the Body. New York: Norton, CJ. Ubiquitous overexpression of CuZn superoxide dismutase 1932. does not extend life span in mice. J Gerontol A Biol Sci Med Sci 13. Caratero A, Courtade M, Bonnet L, Planel H, and Car- 55: B5–B9, 2000. atero C. Effect of a continuous gamma irradiation at a very low 38. Kanungo MS. A model for ageing. J Theor Biol 53: 253–261, dose on the life span of mice. Gerontology 44: 272–276, 1998. 1975. 14. Carey JR. Theories of life span and aging. In: Physiological Basis of Aging and Geriatrics (3rd ed.), edited by Timiras PS. 39. Kawano T, Ito Y, Ishiguro M, Takuwa K, Nakajima T, and Boca Raton, FL: CRC, 2003, p. 85–95. Kimura Y. Molecular cloning and characterization of a new 15. Chang E and Harley CB. Telomere length and replicative insulin/IGF-like peptide of the nematode Caenorhabditis el- aging in human vascular tissues. Proc Natl Acad Sci USA 92: egans. Biochem Biophys Res Commun 273: 431–436, 2000. 11190–11194, 1995. 40. Keller L and Genoud M. Extraordinary lifespans in ants: a 16. Chin L, Artandi SE, Shen Q, Tam A, Lee SL, Gottlieb GJ, test of evolutionary theories of ageing. Nature 389: 958–960, Greider CW, and DePinho RA. p53 Deficiency rescues the 1997. adverse effects of telomere loss and cooperates with telomere 41. Kim NW, Piatyszek MA, Prowse KR, Harley CB, West MD, dysfunction to accelerate carcinogenesis. Cell 97: 527–538, Ho PL, Coviello GM, Wright WE, Weinrich SL, and Shay 1999. JW. Specific association of human telomerase activity with 17. Collins K. Mammalian telomeres and telomerase. Curr Opin immortal cells and cancer. Science 266: 2011–2015, 1994. Cell Biol 12: 378–383, 2000. 42. Kirkwood TB. New science for an old problem. Trends Genet 18. Collins K and Wong JMY. Telomere maintenance and dis- 18: 441–442, 2002. ease. Lancet, Published online May 13, 2003. http://www. 43. Kowald A and Kirkwood TB. A network theory of ageing: the thelancet.com/journal/vol361/iss9370/abs/llan.361.9370.early_ interactions of defective mitochondria, aberrant proteins, free online_publication.25751.1. radicals and scavengers in the ageing process. Mutat Res 316: 19. Daynes RA and Araneo BA. Prevention and reversal of some 209–236, 1996. age-associated changes in immunologic responses by supple- 44. Krtolica A and Campisi J. Cancer and aging: a model for the mental dehydroepiandrosterone sulfate therapy. Aging: Immu- cancer promoting effects of the aging stroma. Int J Biochem Cell nology, and Infectious Disease 3: 135–153, 1992. Biol 34: 1401–1414, 2002. 20. Dimri GP, Lee X, Basile G, Acosta M, Scott G, Roskelley 45. Krtolica A, Parrinello S, Lockett S, Desprez PY, and C, Medrano EE, Linskens M, Rubelj I, Pereira-Smith O, et Campisi J. Senescent fibroblasts promote epithelial cell

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org INVITED REVIEW 1715

growth and tumorigenesis: a link between cancer and aging. duction in Drosophila melanogaster. Proc R Soc Lond B Biol Sci Proc Natl Acad Sci USA 98: 12072–12077, 2001. 266: 255–261, 1999. 46. Larsen PL. Aging and resistance to oxidative damage in Cae- 69. Perls T, Kunkel L, and Puca A. The genetics of aging. Curr norhabditis elegans. Proc Natl Acad Sci USA 90: 8905–8909, Opin Genet Dev 12: 362–369, 2002. 1993. 70. Perls TT, Wilmoth J, Levenson R, Drinkwater M, Cohen 47. Lebel M and Leder P. A deletion within the murine Werner M, Bogan H, Joyce E, Brewster S, Kunkel L, and Puca A. syndrome helicase induces sensitivity to inhibitors of topoisom- Life-long sustained mortality advantage of siblings of centenar- erase and loss of cellular proliferative capacity. Proc Natl Acad ians. Proc Natl Acad Sci USA 99: 8442–8447, 2002. Sci USA 95: 13097–13102, 1998. 71. Pletcher SD, Macdonald SJ, Marguerie R, Certa U, 48. Lee CK, Klopp RG, Weindruch R, and Prolla TA. Gene Stearns SC, Goldstein DB, and Partridge L. Genome-wide expression profile of aging and its retardation by caloric restric- transcript profiles in aging and calorically restricted Drosophila tion. Science 285: 1390–1393, 1999. melanogaster. Curr Biol 12: 712–723, 2002. 49. Lee HW, Blasco MA, Gottlieb GJ, Horner JW 2nd, Greider 72. Puca AA, Daly MJ, Brewster SJ, Matise TC, Barrett J, CW, and DePinho RA. Essential role of mouse telomerase in Shea-Drinkwater M, Kang S, Joyce E, Nicoli J, Benson E, highly proliferative organs. Nature 392: 569–574, 1998. Kunkel LM, and Perls T. A genome-wide scan for linkage to 50. Linskens MH, Feng J, Andrews WH, Enlow BE, Saati SM, human exceptional longevity identifies a locus on chromosome Tonkin LA, Funk WD, and Villeponteau B. Cataloging 4. Proc Natl Acad Sci USA 98: 10505–10508, 2001. altered gene expression in young and senescent cells using 73. Reddel RR. The role of senescence and immortalization in enhanced differential display. Nucleic Acids Res 23: 3244–3251, carcinogenesis. Carcinogenesis 21: 477–484, 2000. 1995. 74. Reinke V. Functional exploration of the C. elegans genome 51. Loison A, Festa-Bianchet M, Gaillard JM, Jorgenson JT, using DNA microarrays. Nat Genet 32, Suppl: 541–546, 2002. and Jullien JM. Age-specific survival in five populations of 75. Rose M and Charlesworth B. A test of evolutionary theories ungulates: evidence of senescence. Ecology 80: 2539–2554, of senescence. Nature 287: 141–142, 1980. 1999. 76. Rowe JW and Kahn RL. Human aging: usual and successful. 52. Maier JA, Voulalas P, Roeder D, and Maciag T. Extension Science 237: 143–149, 1987. of the life-span of human endothelial cells by an interleukin-1 77. Rowe JW and Kahn RL. Successful aging. Aging (Milano) 10: alpha antisense oligomer. Science 249: 1570–1574, 1990. 142–144, 1998. 53. Mandavilli BS, Santos JH, and Van Houten B. Mitochon- 78. Rubin H. The disparity between human cell senescence in vitro drial DNA repair and aging. Mutat Res 509: 127–151, 2002. and lifelong replication in vivo. Nat Biotechnol 20: 675–681, 2002. 54. Martin GM and Oshima J. Lessons from human progeroid 79. Sacher GA. Evolutionary theory in gerontology. Perspect Biol syndromes. Nature 408: 263–266, 2000. Med 25: 339–353, 1982. 55. Martin GM, Sprague CA, and Epstein CJ. Replicative life- 80. Sapolsky RM. Stress, the Aging Brain, and the Mechanisms of Neuron Death. Cambridge, MA: MIT Press, 1992. span of cultivated human cells. Effects of donor’s age, tissue, 81. Sapolsky RM. Why Zebras Don’t Get Ulcers: An Updated and genotype. Lab Invest 23: 86–92, 1970. Guide to Stress, Stress-Related Diseases, and Coping. New 56. Maser RS and DePinho RA. Connecting chromosomes, crisis, York: Freeman, 1998. and cancer. Science 297: 565–569, 2002. 82. Selye H. The Stress of Life. New York: McGraw-Hill, 1976. 57. Masoro EJ and Austad SN. The evolution of the antiaging 83. Sgro CM and Partridge L. A delayed wave of death from action of dietary restriction: a hypothesis. J Gerontol A Biol Sci reproduction in Drosophila. Science 286: 2521–2524, 1999. Med Sci 51: B387–B391, 1996. 84. Sherr CJ and DePinho RA. Cellular senescence: mitotic clock 58. McEwen BS. The End of Stress as We Know It. Washington, or culture shock? Cell 102: 407–410, 2000. DC: Joseph Henry Press, 2002. 85. Shock NW. Age changes in physiological functions in the total 59. Medawar PB. An Unsolved Problem in Biology. London: animal: the role of tissue loss. In: The Biology Of Aging: A Lewis, 1952. Symposium (3rd ed.), edited by Strehler BL, Ebert JD, and 60. Mehlhorn RJ. Oxidants and antioxidants in aging. In: Physi- Shock NW. Washington, DC: Am. Inst. Biol. Sci., 1960. ological Basis of Aging and Geriatrics (3rd ed.), edited by 86. Shringarpure R and Davies KJ. Protein turnover by the Timiras PS. Boca Raton, FL: CRC, 2003, p. 61–83. proteasome in aging and disease. Free Radic Biol Med 32: 61. Melov S, Ravenscroft J, Malik S, Gill MS, Walker DW, 1084–1089, 2002. Clayton PE, Wallace DC, Malfroy B, Doctrow SR, and 87. Tatar M, Bartke A, and Antebi A. The endocrine regulation Lithgow GJ. Extension of life-span with superoxide dis- of aging by insulin-like signals. Science 299: 1346–1351, 2003. mutase/catalase mimetics. Science 289: 1567–1569, 2000. 88. Timiras P. Neuroendocrine models regulating lifespan. In: 62. Merry BJ. Dietary restriction in aging. In: Studies of Aging, Development, Maturation and Senescence of Neuroendocrine edited by Sternberg H and Timiras PS. New York: Springer, Systems: A Comparative Approach, edited by Schreibman MP 1999. and Scanes CG. San Diego, CA: Academic, 1989. 63. Merry BJ. Molecular mechanisms linking calorie restriction 89. Timiras PS. Biological perspectives on aging. Am Sci 66: and longevity. Int J Biochem Cell Biol 34: 1340–1354, 2002. 605–613, 1978. 64. Miller JK. Escaping senescence: demographic data from the 90. Timiras PS. Physiological Basis of Aging and Geriatrics (3rd three-toed box turtle (Terrapene carolina triunguis). Exp Ger- ed.), edited by Timiras PS. Boca Raton, FL: CRC, 2003. ontol 36: 829–832, 2001. 91. Tower J. Transgenic methods for increasing Drosophila life 65. Mitchell JR, Wood E, and Collins K. A telomerase compo- span. Mech Ageing Dev 118: 1–14, 2000. nent is defective in the human disease dyskeratosis congenita. 92. Tyner SD, Venkatachalam S, Choi J, Jones S, Ghebrani- Nature 402: 551–555, 1999. ous N, Igelmann H, Lu X, Soron G, Cooper B, Brayton C, 66. Mobbs CV, Bray GA, Atkinson RL, Bartke A, Finch CE, Hee Park S, Thompson T, Karsenty G, Bradley A, and Maratos-Flier E, Crawley JN, and Nelson JF. Neuroendo- Donehower LA. p53 mutant mice that display early ageing- crine and pharmacological manipulations to assess how caloric associated phenotypes. Nature 415: 45–53, 2002. restriction increases life span. J Gerontol A Biol Sci Med Sci 56, 93. Udelsman R, Blake MJ, Stagg CA, Li DG, Putney DJ, and Spec No 1: 34–44, 2001. Holbrook NJ. Vascular heat shock protein expression in re- 66a.Parrinello S, Samper E, Krtolica A, Goldstein J, Melov S, sponse to stress. Endocrine and autonomic regulation of this and Campisi J. Oxygen sensitivity severely limits the replica- age-dependent response. J Clin Invest 91: 465–473, 1993. tive life span of murine cells. Nat Cell Biol. In press. 94. Vasile E, Tomita Y, Brown LF, Kocher O, and Dvorak HF. 67. Partridge L and Gems D. Mechanisms of ageing: public or Differential expression of thymosin ␤-10 by early passage and private? Nat Rev Genet 3: 165–175, 2002. senescent vascular endothelium is modulated by VPF/VEGF: 68. Partridge L, Prowse N, and Pignatelli P. Another set of evidence for senescent endothelial cells in vivo at sites of ath- responses and correlated responses to selection on age at repro- erosclerosis. FASEB J 15: 458–466, 2001.

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org 1716 INVITED REVIEW

95. Vulliamy T, Marrone A, Goldman F, Dearlove A, Bessler 100. Weindruch R and Walford RL. The Retardation of Aging M, Mason PJ, and Dokal I. The RNA component of telomer- and Diseases by Dietary Restriction. Springfield, IL: Thomas, ase is mutated in autosomal dominant dyskeratosis congenita. 1998. Nature 413: 432–435, 2001. 101. Wilmoth JR. Human longevity in historical perspective. In: 96. Wachter KW and Finch CE. Between Zeus and the Salmon. Physiological Basis of Aging and Geriatrics (3rd ed.), edited by Washington, DC: National Academy Press, 1997. Timiras PS. Boca Raton, FL: CRC, 2003, p. 11–24. 97. Wakayama T, Shinkai Y, Tamashiro KL, Niida H, Blanchard 102. Wright WE, Piatyszek MA, Rainey WE, Byrd W, and Shay DC, Blanchard RJ, Ogura A, Tanemura K, Tachibana M, JW. Telomerase activity in human germline and embryonic Perry AC, Colgan DF, Mombaerts P, and Yanagimachi R. tissues and cells. Dev Genet 18: 173–179, 1996. Cloning of mice to six generations. Nature 407: 318–319, 2000. 103. Wright WE and Shay JW. Historical claims and current 98. Weindruch R, Kayo T, Lee CK, and Prolla TA. Gene ex- interpretations of replicative aging. Nat Biotechnol 20: 682– pression profiling of aging using DNA microarrays. Mech Age- 688, 2002. ing Dev 123: 177–193, 2002. 104. Zou S, Meadows S, Sharp L, Jan LY, and Jan YN. Ge- 99. Weindruch R, Kayo T, Lee CK, and Prolla TA. Microarray nome-wide study of aging and oxidative stress response in profiling of gene expression in aging and its alteration by caloric Drosophila melanogaster. Proc Natl Acad Sci USA 97: restriction in mice. J Nutr 131: 918S–923S, 2001. 13726–13731, 2000.

J Appl Physiol • VOL 95 • OCTOBER 2003 • www.jap.org